Forensic Review of "β-Ecdysone-Loaded Gelatin Methacryloyl Hydrogel Accelerates Diabetic Wound Healing by Improving Angiogenesis" (ACS Omega, DOI: 10.1021/acsomega.5c11972)

Academic fraud report · Geng Detector

Summary

Verdict: No clear indicators of academic misconduct detected based on the available textual content. The paper, authored by Ziwei Wu and colleagues and published in ACS Omega, was reviewed across multiple dimensions. Key checks included: (1) internal consistency of physicochemical data for the hydrogel characterization, which showed plausible SD variability and a coherent logical pattern (β-E, a hydrophobic small molecule, reducing swelling and slowing degradation of GelMA); (2) a coherent publication timeline (submitted November 2025, accepted April 2026) with up-to-date citations, including 2026 references; (3) rigorous and unusually detailed statistical methodology (normality and log-normality testing prior to one-way ANOVA, n=3 for in vitro assays, 24 rats in 4 groups); and (4) no signs of fabricated data patterns such as suspiciously uniform SDs. Limitations: only textual content was analyzed; no raw image files were available, so pixel-level checks (image reuse, splicing/tampering) could not be performed. Confidence in the textual analysis is moderate-to-high; image-based concerns remain unaddressed.

Verdict

No substantive evidence of academic misconduct was identified from the available text. Image-level forensic checks (duplicate-image detection, splicing/Photoshop artifacts) could not be performed because only the manuscript text was provided. Based on text alone, the paper appears internally consistent, methodologically rigorous, and temporally coherent.

Key findings

Hydrogel characterization data are self-consistent. Swelling ratio: β-E/GelMA = 372.07 ± 9.54% vs GelMA = 451.95 ± 28.56%; compressive modulus: GelMA = 16.06 ± 3.36 kPa vs β-E/GelMA = 17.10 ± 2.00 kPa; degradation: β-E/GelMA = 58.57 ± 1.62% vs GelMA = 61.54 ± 0.95%. SDs are uneven and within plausible experimental noise.
Logical closure on physicochemistry. Loading a hydrophobic small molecule (β-ecdysone) into GelMA plausibly reduces swelling ratio and slows degradation, providing a mechanistically coherent explanation for the observed trends — a pattern inconsistent with synthetic "random number generator" data fabrication.
Timeline is internally coherent. Submitted 2025-11-13, accepted 2026-04-07. References include 2024–2026 publications (e.g., Refs. 10, 29, 49 in 2026 issues), consistent with a draft prepared at submission time. No anachronistic reagents or methods were detected.
Statistical methodology is unusually explicit. Authors used GraphPad Prism 9.5.1, performed normality and log-normality tests prior to one-way ANOVA — uncommon in papers flagged for fabrication. Sample sizes (n = 3 for in vitro; 24 rats across 4 groups) are conventional.
Minor typographical artifact in Section 3.4: "intheConβ-E/GelMAgroups" appears to be a PDF text-extraction artifact rather than a substantive issue.

Evidence highlights

Swelling ratio differential of ~80 percentage points (372.07 vs 451.95) with non-identical SDs (9.54 vs 28.56) — consistent with real experimental variability.
Compressive modulus difference of ~1 kPa (16.06 vs 17.10) with overlapping but distinct SDs (3.36 vs 2.00) — typical of biomaterial mechanical testing.
Degradation difference of ~3 percentage points (58.57 vs 61.54) at day-equivalent time points with small SDs (~1%) — plausible for enzymatic or hydrolytic degradation assays.
Statistical workflow explicitly states normality/log-normality testing before ANOVA — a marker rarely seen in fabrication cases.
DOI: 10.1021/acsomega.5c11972.

Notes

Critical limitation: Figures 1–5 (and subpanels) could not be examined at the pixel level. The reviewer flagged this as an informational gap, not as a positive finding.
The "Geng six-style" image-reuse (式1) and image-splicing (式3) checks remain pending pending access to original high-resolution figure files.
A minor PDF text-extraction artifact ("intheConβ-E/GelMAgroups") is noted but judged to be a rendering/PDF parsing issue, not a content or integrity issue.
Confidence: moderate-to-high for text-based analyses; low-to-none for image integrity. Any future image acquisition should trigger re-review under 式1 and 式3.